Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A storage device comprising: a first nonvolatile memory set (NVM set) including at least one first namespace; a second NVM set including at least one second namespace, the second NVM set physically isolated from the first NVM set; and a storage controller configured to perform an operation in one of the first and second NVM sets based on which of a first submission queue and a second submission queue of an external device a command is queued, wherein the storage controller is configured to map the first and second submission queues respectively to the first and second NVM sets, perform in the first NVM set a first operation corresponding to a first command fetched from the first submission queue, and perform in the second NVM set a second operation corresponding to a second command fetched from the second submission queue.
A storage device includes multiple nonvolatile memory (NVM) sets that are physically isolated from each other, each containing at least one namespace. The device also includes a storage controller that processes commands based on which submission queue of an external device they originate from. The controller maps specific submission queues to specific NVM sets, ensuring that commands from one queue are executed only in the corresponding NVM set. For example, a first command from a first submission queue is processed in a first NVM set, while a second command from a second submission queue is processed in a second NVM set. This isolation prevents interference between operations, improving security and performance by ensuring that data and operations from different sources remain segregated. The physical separation of NVM sets further enhances reliability and fault tolerance, as issues in one set do not affect the others. This design is particularly useful in multi-tenant or high-security environments where strict data isolation is required.
2. The storage device of claim 1 , wherein the first command comprises a first logical address corresponding to the at least one first namespace, and the second command comprises a second logical address corresponding to the at least one second namespace.
A storage device includes a controller and a memory configured to store data in multiple namespaces. The controller is designed to execute commands that access data within these namespaces. The device receives a first command targeting a first namespace and a second command targeting a second namespace. The first command includes a first logical address that maps to the first namespace, while the second command includes a second logical address that maps to the second namespace. The controller processes these commands to read or write data in the respective namespaces based on the provided logical addresses. This allows the storage device to manage and access data across different namespaces efficiently, ensuring proper isolation and organization of data within the memory. The logical addresses serve as identifiers that direct the controller to the correct namespace for data operations, enabling structured data management in a multi-namespace storage environment.
3. The storage device of claim 1 , wherein the storage controller is configured to provide a first Virtual Function (VF) that controls the first NVM set, and a second Virtual Function (VF) that controls the second NVM set.
This invention relates to storage devices with non-volatile memory (NVM) management, specifically addressing the need for efficient and isolated control of multiple NVM sets within a single storage device. The storage device includes a storage controller and at least two distinct NVM sets, each comprising multiple NVM devices. The storage controller is configured to provide separate Virtual Functions (VFs) for each NVM set, enabling independent control and management. A first Virtual Function (VF) is assigned to control the first NVM set, while a second Virtual Function (VF) is assigned to control the second NVM set. This separation allows for isolated operation, improving performance, security, and resource allocation. The storage controller may also include a physical function (PF) that manages the overall operation of the storage device, including the assignment and configuration of the VFs. The use of VFs ensures that each NVM set operates independently, reducing interference and enhancing scalability. This design is particularly useful in environments requiring multi-tenancy or workload isolation, such as cloud computing or enterprise storage systems. The invention improves efficiency by allowing parallel access to different NVM sets while maintaining strict isolation between them.
4. The storage device of claim 1 , wherein the storage controller comprises: a first nonvolatile memory express (NVMe) controller configured to control the first NVM set, based on the first command; a second NVMe controller configured to control the second NVM set, based on the second command; and a peripheral component interconnect express (PCIe) port configured to provide communication between the external device and each of the first and second NVMe controllers.
A storage device includes a storage controller that manages multiple nonvolatile memory (NVM) sets. The storage controller contains two separate NVMe (Nonvolatile Memory Express) controllers, each responsible for controlling a distinct NVM set. The first NVMe controller processes commands directed to the first NVM set, while the second NVMe controller handles commands for the second NVM set. The storage controller also includes a PCIe (Peripheral Component Interconnect Express) port that facilitates communication between an external device and both NVMe controllers. This architecture allows the storage device to independently manage multiple NVM sets, improving performance and flexibility by enabling parallel command processing and efficient data transfer through the PCIe interface. The design ensures that each NVM set operates under its own NVMe controller, reducing bottlenecks and enhancing scalability for high-performance storage applications.
5. The storage device of claim 1 , wherein the first NVM set communicates with the storage controller through at least one first channel and the second NVM set communicates with the storage controller through at least one second channel physically isolated from the at least one first channel.
This invention relates to a storage device with improved data reliability and performance by using physically isolated communication channels between non-volatile memory (NVM) sets and a storage controller. The device includes at least two NVM sets, each connected to the storage controller through separate, physically isolated channels. The first NVM set communicates via at least one first channel, while the second NVM set communicates via at least one second channel that is physically isolated from the first channel. This isolation prevents interference or data corruption between channels, enhancing reliability and performance. The storage controller manages data operations across the NVM sets, ensuring efficient data storage and retrieval. The isolated channels may be implemented using separate physical links, such as dedicated buses or cables, to ensure no shared electrical or signal pathways exist between the channels. This design is particularly useful in high-reliability applications where data integrity and performance are critical, such as enterprise storage systems or mission-critical computing environments. The invention improves upon traditional storage devices that rely on shared communication channels, which can introduce latency, interference, or failure risks.
6. The storage device of claim 1 , wherein the first and second NVM sets are recognized as physically isolated storage areas by the external device.
A storage device includes multiple non-volatile memory (NVM) sets, where each set is recognized by an external device as a physically isolated storage area. The device ensures that data stored in one NVM set is inaccessible to operations targeting another NVM set, providing logical and physical separation. This isolation prevents unauthorized access or interference between different storage areas, enhancing security and reliability. The storage device may include a controller that manages data distribution across the NVM sets, ensuring that each set operates independently while maintaining overall system performance. The isolation can be enforced through hardware partitioning, firmware controls, or a combination of both, depending on the implementation. This design is particularly useful in systems requiring strict data segregation, such as multi-tenant environments, secure storage applications, or devices handling sensitive information. The storage device may also support dynamic allocation of NVM sets based on usage patterns or security policies, allowing flexible adaptation to changing requirements. By maintaining physical isolation, the device ensures that even if one NVM set is compromised, the integrity of data in other sets remains protected.
7. The storage device of claim 1 , wherein the storage controller is configured to write first completion information of the first operation corresponding to the first command in a first completion queue of the external device, after completing the first operation, and the storage controller is configured to write second completion information of the second operation corresponding to the second command in a second completion queue of the external device, after completing the second operation.
A storage device includes a storage controller that processes commands from an external device, such as a host system. The storage controller executes operations in response to received commands and generates completion information upon finishing each operation. The completion information indicates the status of the completed operation, such as success or failure. The storage controller writes this completion information into designated completion queues within the external device. Specifically, after completing a first operation corresponding to a first command, the storage controller writes first completion information into a first completion queue of the external device. Similarly, after completing a second operation corresponding to a second command, the storage controller writes second completion information into a second completion queue of the external device. This mechanism allows the external device to track the status of multiple operations independently, improving efficiency and reducing the need for repeated status checks. The storage controller may also manage multiple command queues and completion queues, ensuring proper synchronization between operations and their corresponding completion notifications. This system enhances performance by minimizing latency and ensuring reliable communication between the storage device and the external device.
8. An operation method of a storage device comprising a first nonvolatile memory set (NVM set) and a second NVM set, the method comprising: fetching by the storage device a command from an external device comprising first and second submission queues; and performing an operation in one of the first and second NVM sets based on which of the first and second submission queues the command is queued, wherein said performing comprises mapping the first and second submission queues respectively to the first and second NVM sets, performing by the storage device a first operation corresponding to the command in the first NVM set when the command is fetched from the first submission queue, and then transmitting by the storage device first completion information of the first operation to the external device, and performing by the storage device a second operation corresponding to the command in the second NVM set when the command is fetched from the second submission queue, and then transmitting by the storage device second completion information of the second operation to the external device.
This invention relates to storage devices with multiple nonvolatile memory (NVM) sets and a method for managing command processing from external devices. The problem addressed is efficient command handling in storage systems where multiple NVM sets are present, ensuring proper routing and completion reporting. The method involves a storage device with at least two NVM sets and an external device that submits commands via two separate submission queues. The storage device fetches commands from these queues and performs operations in the corresponding NVM set based on the queue from which the command was fetched. Commands from the first submission queue are processed in the first NVM set, with completion information sent back to the external device upon operation completion. Similarly, commands from the second submission queue are processed in the second NVM set, with separate completion information transmitted. This approach ensures that commands are routed to the correct NVM set and completion status is accurately reported, improving efficiency and reliability in multi-NVM storage systems. The method avoids contention between operations by isolating command processing paths based on queue assignment.
9. The method of claim 8 , wherein the first and second NVM sets are connected with different physical channels, and wherein each of the first and second NVM sets includes at least one nonvolatile memory device connected with the different physical channels.
This invention relates to a method for managing data storage in a system with multiple nonvolatile memory (NVM) sets. The problem addressed is improving data reliability and performance by distributing storage operations across physically separate channels. The method involves using at least two NVM sets, each connected to different physical channels, where each set includes one or more nonvolatile memory devices. By distributing data across these independent channels, the system can enhance fault tolerance, reduce bottlenecks, and improve overall storage efficiency. The different physical channels ensure that failures in one channel do not affect the other, providing redundancy and resilience. This approach is particularly useful in high-availability storage systems where data integrity and performance are critical. The method may be applied in various storage architectures, including solid-state drives (SSDs) and enterprise storage solutions, to optimize data access and minimize downtime. The use of separate channels also allows for parallel operations, further improving throughput and reducing latency.
10. The method of claim 8 , wherein the first completion information is written in a first completion queue of the external device, and the second completion information is written in a second completion queue of the external device.
This invention relates to data processing systems, specifically methods for managing completion information in external devices. The problem addressed is the efficient handling of completion information generated by external devices, such as storage controllers or network adapters, to ensure proper synchronization and processing of tasks in a computing system. The method involves writing first completion information to a first completion queue in an external device and second completion information to a second completion queue in the same external device. The external device may be a storage controller, network adapter, or other peripheral device that generates completion information upon finishing tasks. The first and second completion queues are distinct, allowing for separate handling of different types of completion events or prioritization of tasks. This separation improves system performance by reducing contention and ensuring timely processing of critical operations. The method may also include steps to monitor the queues, retrieve completion information, and update system state based on the retrieved data. The use of multiple queues enhances scalability and reliability in high-performance computing environments.
11. A storage device comprising: a first nonvolatile memory set (NVM set) including at least one first namespace; a second NVM set including at least one second namespace, the second NVM set physically isolated from the first NVM set; and a storage controller configured to fetch a command from an external device, select one of the first and second NVM sets, and perform an operation corresponding to the fetched command in the one of the first and second NVM sets, wherein the storage controller is configured to map first and second submission queues from among a plurality of submission queues respectively to the first and second NVM sets, and select the one of the first and second NVM sets depending on which of the first and second submission queues the fetched command is fetched.
This invention relates to a storage device with physically isolated nonvolatile memory (NVM) sets to enhance data security and performance. The device includes at least two NVM sets, each containing one or more namespaces, where the sets are physically isolated to prevent unauthorized access or interference between them. A storage controller manages operations by fetching commands from an external device and selecting the appropriate NVM set based on the command's origin. The controller maps specific submission queues to each NVM set, ensuring commands from different queues are directed to their designated NVM set. This isolation allows for secure data partitioning, where sensitive data can be stored in one NVM set while general data resides in another, reducing the risk of cross-contamination or unauthorized access. The system improves performance by reducing contention between different types of workloads, as commands are routed to the correct NVM set without unnecessary delays. The storage controller's ability to dynamically select the appropriate NVM set based on submission queue mappings ensures efficient and secure data handling. This design is particularly useful in environments requiring high security, such as enterprise storage systems or multi-tenant cloud storage, where data isolation is critical.
Unknown
October 27, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.